During testing, it was discovered that when a device tries to attach to the
alua handler while in TPG state of transitioning, the alua_rtpg function will
wait for it to exit the state before allowing it to continue. As a result, if
the 60 second timeout expires, the alua handler will not attach to the device.

To fix this, I have introduced an input argument to alua_rtpg called
wait_for_transition.  The idea is that it will wait for the transition to
complete before an activation (because the current TPG state has some bearing
in that case), but during a discovery if it is transitioning, it will not
wait, and will store the state as standby for the time being.

I believe the precedent exists for this from commit
c0d289b3 Since if the device reports a state
of transitioning, it can transition to other more valid states, and it has
been established TPGS is supported on the device, if it is attaching.
Signed-off-by: NSean Stewart <Sean.Stewart@netapp.com>
Acked-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

When the scsi_dh_alua handler issues an RTPG during initialization, if it gets
0x06/0x29/0x04 as the sense, it will fail to attach the handler.  NetApp
E-Series returns 0x29/0x00 for power on, and 0x29/0x04 for conditions that
cause the controller to reboot again.  These conditions should be treated
identically within the handler.
Signed-off-by: NSean Stewart <Sean.Stewart@netapp.com>
Acked-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Resending with requested rewording of the parameter description.  Currently
ALUA device handler sends STPG command during failover and failback.  Failover
can be optimized by implicit failover (by not to sending STPG command), when 1
is passed as hwhandler parameter in multipath.conf. ex "2 alua 1". We may need
to pass the parameter through module param for alua device handler to optimize
failover if incase retain_attached_hwhandler set in multipath.conf and
hwhandler is set with non-tpgs device handler ex: '1 rdac'.

[jejb: fix up whitespace and other issues]
Signed-off-by: NVijay Chauhan <vijay.chauhan@netapp.com>
Signed-off-by: NSean Stewart <Sean.Stewart@netapp.com>
Acked-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

For the stpg_endio path we are not evaluating the sense. The bug
is that

1. The error value is set to -EIO when there is sense, so we hit the first
error check and always return SCSI_DH_IO.

2. h->senselen is set to zero in submit_stpg. It is not later set to
req->sense_len like in the synchrounous exection paths, so we must
check the req->sense_len field.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

A quote from SPC-4: "While in the unavailable primary target port
asymmetric access state, the device server shall support those of
the following commands that it supports while in the active/optimized
state: [ ... ] d) SET TARGET PORT GROUPS; [ ... ]". Hence enable
sending STPG to a target port group that is in the unavailable state.
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>
Acked-by: NHannes Reinecke <hare@suse.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

A quote from SPC-4: "While in the unavailable primary target port
asymmetric access state, the device server shall support those of
the following commands that it supports while in the active/optimized
state: [ ... ] d) SET TARGET PORT GROUPS; [ ... ]". Hence re-enable
sending STPG to a target port group that is in the unavailable state.
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Reviewed-by: Babu Moger <babu.moger@netapp.com> 
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Currently the backoff algorithm for when to retry alua rtpg
requests progresses geometrically as so:

2, 4, 8, 16, 32, 64... seconds.

This progression can lead to un-needed delay in retrying
alua rtpg requests when the rtpgs are delayed.  A less
aggressive backoff algorithm that is additive would not
lead to such large jumps when delays start getting long, but
would backoff linearly:

2, 4, 6, 8, 10... seconds.
Signed-off-by: NMartin George <marting@netapp.com>
Signed-off-by: NRob Evers <revers@redhat.com>
Reviewed-by: NBabu Moger <babu.moger@netapp.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Some storage arrays are known to return 'illegal request'
when an rtpg extended header request is made.  T10 says the
array should ignore the bit, and return the non-extended
rtpg as the array doesn't support the request.  Working
around this by retrying the rtpg request without the extended
header bit set when the extended rtpg request results in
illegal request.
Signed-off-by: NRob Evers <revers@redhat.com>
Reviewed-by: NBabu Moger <babu.moger@netapp.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

During alua transitions, an array can return transitioning
status in response to rtpg requests.  These requests get
retried for a maximum of 60 seconds by default before timing
out.  Sometimes this timeout isn't sufficient to allow the
array to complete the transition.  T10-spc4 addresses this
under 'Report Target Port Groups' command.

This update retrieves the timeout value from the storage
array if available and retries the transitioning rtpgs
for up to the 'implied transitioning timeout' value
Signed-off-by: NRob Evers <revers@redhat.com>
Reviewed-by: NBabu Moger <babu.moger@netapp.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

This patch optimizes the set target port group(STPG) command. During our
testing, we found that it is not optimal to send stpg command every time
the path group switch happens. This patch uses PREF (preferred target port)
bit with combination of flags passed by multipath user level tool to
optimize this behaviour. If PREF bit is set then it issues a STPG command,
otherwise it will let implicit transfer take place.

By default there is no change in the behaviour. User tool needs to pass the
parameter to make this change take effect. Patch has been tested on NetApp
E series storage.
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

PREF bit indicates preferred target port group for accessing a logical
unit. This bit is used to optimize the STPG command handling.
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Handler expects only one parameter to set the flag ALUA_OPTIMIZE_STPG.
This flag is used to optimize the STPG behaviour. There is no change in
behaviour by default.

For example, to set the flag pass the following parameters from multipath.conf
hardware_handler        "2 alua 1"
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

This patch adds a check-condition in scsi_dh_alua handler for a retry.
Sometimes, I have seen attach failing due to this check-condition with
following error messages on NetApp E series storage.

Dec  7 15:31:01 nilgiris kernel: [102979.696673] scsi 3:0:2:9: alua: port group 00 rel port 01
Dec  7 15:31:01 nilgiris kernel: [102979.697082] scsi 3:0:2:9: alua: rtpg failed with 8000002
Dec  7 15:31:01 nilgiris kernel: [102979.697086] scsi 3:0:2:9: alua: rtpg sense code 06/2a/01
Dec  7 15:31:01 nilgiris kernel: [102979.697088] scsi 3:0:2:9: alua: not attached
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

This patch corrects the retry interval for alua rtpg command. Purpose was to retry the commands in seconds.
But that was not happening. Reason is msleep takes argument in milliseconds.

Also added minor text after successful attach.
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Acked-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

The module.h header was implicitly present everywhere, so files
with no explicit include of the module infrastructure would build
anyway.  We are now removing the implicit include, and so we need
to call out the module.h file that we need explicitly.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

This patch adds one more check-condition for scsi_dh_alua handler. Without
this, the handler attach fails sometimes during the discovery. I have noticed
this with NetApp E-Series storage with alua mode. Also removed some
unnecessary brackets {} for consistency.
Signed-off-by: NBabu Moger <babu.moger@netapp.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

The alua device handler starts the first retry after 10 seconds,
and increases it times 10 for each round.
This leads to an unnecessary delay. This patch modifies it to
start after one second, and increase by a factor of two.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

For Target Portal Group IDs occupying the full 2 bytes in the
RTPG response, the following group_id check in the alua_rtpg
routine always fails in scsi_dh_alua.c:

if (h->group_id == (ucp[2] << 8) + ucp[3]) {

This causes the ALUA handler to wrongly identify the AAS of
a specified device as well as incorrectly interpreting the
supported AAS of the target as seen by the following entries
in the /var/log/messages:

"alua: port group 3ea state A supports tousna"
"alua: port group 3e9 state A supports tousna"

This is because 'ucp' is wrongly declared in alua_rtpg as
a character pointer instead of an unsigned character pointer.
Signed-off-by: NMartin George <marting@netapp.com>
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

When activating a patch we should always update the TPGS state
as it might have changed in between.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Some device handler types are not tied to the vendor/model
but rather to a specific capability. Eg ALUA is supported
if the 'TPGS' setting in the standard inquiry is set.
This patch implements a 'match' callback for device handler
which supersedes the original vendor/model lookup and
implements the callback for the ALUA handler.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Instead of issuing a standard inquiry from within the
alua device handler we can evaluate the TPGS setting from
the existing inquiry data of the sdev and save us the I/O.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

The SCSI ALUA handler currently fails to attach to devices
reporting an UNAVAILABLE/OFFLINE AAS. But given that an
UNAVAILABLE/OFFLINE AAS can transition to other states
like ACTIVE/OPTIMIZED, ACTIVE/NON-OPTIMIZED, etc. as per
SPC4, this ALUA handler behavior should be rectified so
as to attach to devices which also report an
UNAVAILABLE/OFFLINE AAS.
Signed-off-by: NMartin George <marting@netapp.com>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: NLucas De Marchi <lucas.demarchi@profusion.mobi>

instead of doing sizeof(struct X) it's better to do sizeof(*v) where v
is the variable pointing to struct X.
Signed-off-by: NHillf Danton <dhillf@gmail.com>
Signed-off-by: NBabu Moger <babu.moger@lsi.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Currently NetApp's VID/PID details in the INQUIRY response shows up as
'NETAPP' and 'LUN'.  With upcoming scalable SAN ONTAP version on NetApp
controllers, the PID entry alone is being modified to 'LUN C-Mode' (to
distinguish current ONTAP LUNs from scalable ONTAP LUNs).

'LUN' would still suffice for matching 'LUN C-Mode' but best to
explicitly add these new NetApp LUNs to the device list.
Reported-by: NMartin George <marting@netapp.com>
Acked-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Adds Promise VTrak devices to the ALUA device handler.
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NJoseph Gruher <joseph.r.gruher@intel.com>
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Initialize stpg_endio() 'err' to SCSI_DH_OK and only change it to
SCSI_DH_IO accordingly.  This allows the switching of target group state
to be properly reported when no error has occurred.
Signed-off-by: NJoseph Gruher <joseph.r.gruher@intel.com>
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

The use of blk_execute_rq_nowait() implies __blk_put_request() is needed
in stpg_endio() rather than blk_put_request() -- blk_finish_request() is
called with queue lock already held.
Signed-off-by: NJoseph Gruher <joseph.r.gruher@intel.com>
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
Cc: stable@kernel.org
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

submit_stpg() will always return failure so alua_activate() will report
failure via dm-multipath callback function.  Even though the stpg fired
successfuly dm-multipath does not know and always fails to change the
valid path.

By returning SCSI_DH_OK we're now skipping alua_activate()'s call to
activate_complete 'fn'.  But this is fine because stpg_endio() will call
it via h->callback_fn().
Signed-off-by: NJoseph Gruher <joseph.r.gruher@intel.com>
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Snitzer <snitzer@redhat.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

For ALUA we should be handling all states, independent of whether
the mode is explicit or implicit. For 'Transitioning' we should retry
for a certain amount of time; after that we're setting the port
to 'Standby' and return SCSI_DH_RETRY to signal upper layers
a retry is in order here.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Acked-by: NMike Snitzer <snitzer@redhat.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h

percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files.  percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.

percpu.h -> slab.h dependency is about to be removed.  Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability.  As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.

  http://userweb.kernel.org/~tj/misc/slabh-sweep.py

The script does the followings.

* Scan files for gfp and slab usages and update includes such that
  only the necessary includes are there.  ie. if only gfp is used,
  gfp.h, if slab is used, slab.h.

* When the script inserts a new include, it looks at the include
  blocks and try to put the new include such that its order conforms
  to its surrounding.  It's put in the include block which contains
  core kernel includes, in the same order that the rest are ordered -
  alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
  doesn't seem to be any matching order.

* If the script can't find a place to put a new include (mostly
  because the file doesn't have fitting include block), it prints out
  an error message indicating which .h file needs to be added to the
  file.

The conversion was done in the following steps.

1. The initial automatic conversion of all .c files updated slightly
   over 4000 files, deleting around 700 includes and adding ~480 gfp.h
   and ~3000 slab.h inclusions.  The script emitted errors for ~400
   files.

2. Each error was manually checked.  Some didn't need the inclusion,
   some needed manual addition while adding it to implementation .h or
   embedding .c file was more appropriate for others.  This step added
   inclusions to around 150 files.

3. The script was run again and the output was compared to the edits
   from #2 to make sure no file was left behind.

4. Several build tests were done and a couple of problems were fixed.
   e.g. lib/decompress_*.c used malloc/free() wrappers around slab
   APIs requiring slab.h to be added manually.

5. The script was run on all .h files but without automatically
   editing them as sprinkling gfp.h and slab.h inclusions around .h
   files could easily lead to inclusion dependency hell.  Most gfp.h
   inclusion directives were ignored as stuff from gfp.h was usually
   wildly available and often used in preprocessor macros.  Each
   slab.h inclusion directive was examined and added manually as
   necessary.

6. percpu.h was updated not to include slab.h.

7. Build test were done on the following configurations and failures
   were fixed.  CONFIG_GCOV_KERNEL was turned off for all tests (as my
   distributed build env didn't work with gcov compiles) and a few
   more options had to be turned off depending on archs to make things
   build (like ipr on powerpc/64 which failed due to missing writeq).

   * x86 and x86_64 UP and SMP allmodconfig and a custom test config.
   * powerpc and powerpc64 SMP allmodconfig
   * sparc and sparc64 SMP allmodconfig
   * ia64 SMP allmodconfig
   * s390 SMP allmodconfig
   * alpha SMP allmodconfig
   * um on x86_64 SMP allmodconfig

8. percpu.h modifications were reverted so that it could be applied as
   a separate patch and serve as bisection point.

Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: NTejun Heo <tj@kernel.org>
Guess-its-ok-by: NChristoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>

Adds IBM Power Virtual SCSI ALUA devices to the ALUA device handler.
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Newer Netapp target software supports ALUA, so
this patch adds them to the scsi_dev_alua dev list.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Make the activate function asynchronous by using blk_execute_rq_nowait()
Signed-off-by: NChandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

Make scsi_dh_activate() function asynchronous, by taking in two additional
parameters, one is the callback function and the other is the data to call
the callback function with.
Signed-off-by: NChandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

alua_activate only sends a STPG if only explicit is suppored.
As a result, for EMC targets that support both we end up doing
a implicit failover when X commands are finally sent to
the other SP.

This patch does a AND on the h->tpgs, so we do a explicit failover
right away.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: NJames Bottomley <James.Bottomley@suse.de>

When we switch controllers the Intel Multi-Flex reports
REPORTED_LUNS_DATA_HAS_CHANGED. This patch just has us
retry the command.
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This adds the Intel Multi-Flex device to scsi_dh_alua's
scsi_dh_devlist, so the module attaches to these devs.
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

The buf[i] is a byte but we are only asking 4 bits off the
group_id. This patch has us take off a byte.
Signed-off-by: NIlgu Hong <ilgu.hong@promise.com>
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

We do not need to set REQ_NOMERGE because when the module calls
blk_execute_rq -> blk_execute_rq_nowait, blk_execute_rq_nowait sets
it for us. This brings all the modules in sync for those bits.
Signed-off-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>